Multistage Computational Evaluation of Bedaquiline Pyridyl Derivatives: Unveiling Anti‐Tubercular Potential Through QSAR, Molecular Docking, Molecular Dynamics, ADMET, and DFT Analysis

ABSTRACT ATP synthase is essential for cellular energy production, and small‐molecule inhibitors of this enzyme have shown significant promise as anti‐tubercular agents. In this study, QSAR analysis was performed using QSARINS to identify potential ATP synthase inhibitors. Among the four developed models, the best‐performing model incorporated seven key descriptors: GATSe3, MinAbsEStateIndex, PEOE_VSA3, MATS7c, VR2_Dzi, VR1_Dzs, and SpMin7_Bhs. The model exhibited strong statistical performance ( R 2 = 0.8518, R 2 adj = 0.8245, RMSE tr = 0.2380, MAE tr = 0.1717, Q 2 LOO = 0.8025, Q 2 F 3 = 0.8097, CCC ext = 0.8798), confirming its robustness and predictive reliability. Based on the validated QSAR model, a series of novel compounds (BP1a–BP1e, BP2a–BP2e, BP3a–BP3e) were designed, and their pMIC 90 values were computationally predicted. BP‐1d emerged as the most promising inhibitor. Molecular docking also showed that the compound BP‐1d with ATP synthase protein had a strong binding affinity (–7.3 kcal/mol), which was better than the reference drug bedaquiline (–6.6 kcal/mol). Molecular dynamics simulations (100 ns) further indicated that BP‐1d forms a more stable enzyme‐ligand complex, maintaining lower RMSD (below 0.5 nm) and consistent hydrogen bonding compared to bedaquiline. Overall, BP‐1d demonstrates strong potential as a lead anti‐tubercular candidate targeting ATP synthase protein.